Species shockiness: first investigations

Cover that shocks which I ought not to see.

1 Pinus ponderosa (PIPO)

1.1 Posterior quantification

diagnostics <- readRDS(file.path(wd, 'model/shocks/output', 'diagnostics_03dec2025_PIPO.rds'))
util$check_all_hmc_diagnostics(diagnostics)

  All Hamiltonian Monte Carlo diagnostics are consistent with reliable
Markov chain Monte Carlo.

1.2 Retrodictive checks

Let’s look at 8 random trees!

2 Pinus edulis (PIED)

2.1 Posterior quantification

diagnostics <- readRDS(file.path(wd, 'model/shocks/output', 'diagnostics_03dec2025_PIED.rds'))
util$check_all_hmc_diagnostics(diagnostics)

  Chain 3: 1024 of 1024 transitions (100.00%)
           saturated the maximum treedepth of 10.

  Numerical trajectories that saturate the maximum treedepth have
terminated prematurely.  Increasing max_depth above 10 should result in
more expensive, but more efficient, Hamiltonian transitions.

2.2 Retrodictive checks

Let’s look at 8 random trees!

3 Pseudotsuga menziesii (PSME)

3.1 Posterior quantification

diagnostics <- readRDS(file.path(wd, 'model/shocks/output', 'diagnostics_03dec2025_PSME.rds'))
util$check_all_hmc_diagnostics(diagnostics)

  All Hamiltonian Monte Carlo diagnostics are consistent with reliable
Markov chain Monte Carlo.

3.2 Retrodictive checks

Let’s look at 8 random trees!

4 Comparison of the 3 species

4.1 Parameters

4.2 Overall shockiness

4.2.1 PIPO

4.2.2 PIED

4.2.3 PSME

4.3 Common stands

4.3.1 PIPO and PIED

No stands in common. But we could try to identify close stands.

S74 and S45 could be a good example. The shock probabilities are not really different:

What about the actual shock behavior? Let’s take 10 random trees on each stand:

4.3.2 PIPO and PSME

For those 2 species, it’s easier because we actually have stands where both were cored. Some of them are different, some of them are similar:

In particular, this stand is interesting:

How to interpret this? On this stand, the shocks are more frequent for PIPO, but less trees experience shocks given a stand is shocking? Or should we rather interpret \phi_\text{shock} \times \omega_\text{shock}^{concordant}?

We can look at 20 trees from each species: